Aliasing is simply caused by the input frequency being greater than 1/2
the sampling frequency. If there were a single frequency present that is
too high it would be reconstructed as a single lower frequency. It's
pretty hard for me to imagine what grain which is too small to be
correctly sampled by the sensor spacing would be reconstructed as ...
but I think what I'm seeing is the grain rather than sampling aritfacts.
The scanner groups seem to have a lot of talk of 2800 ppi being a bigger
problem for grain aliasing than 4000 ppi. I don't know since I haven't
had my scanner very long nor have I tried to compare grain size to
sample spacing.
-jeff
(if the grain lands on multiple pixels it's frequency is lower than if
it is completely contained by a single pixel ... but probably grain that
is always small enough to be completely contained by a single pixel is
effectively low pass filtered unless it is repetitively positioned ...
wild guess on my part).
----- Original Message -----
From: "Andrew Dacey" <frugal@xxxxxxxxxxxxxxx>
>
> ----- Original Message -----
> From: "Jeff Keller" <jrk_om@xxxxxxxxxxx>
>
> > I wouldn't entirely agree with this ... although not entirely
obvious,
> > at 4800 samples per inch, there's something there that looks like
grain
> > to me. The ability to pull detail out of an image is awesome though.
>
> I know one of the concerns with scanning film is "grain aliasing". If
memory
> serves, this ocurs when the resolution of the scan is not high enough
to
> fully resolve the grain. What happens is that grain which does not
land on a
> single pixel gets recorded on more than 1 pixel which can result in
the
> apearance of grain being increased. I'd suspect that at 4800 ppi this
is
> starting to happen and if a greater resolution were used the grain
would be
> fully captured but be less aparent.
>
> Andrew "frugal" Dacey
< This message was delivered via the Olympus Mailing List >
< For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
< Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >
|